A fast algorithm for computing longest common subsequences
Communications of the ACM
Realistic BGP traffic for test labs
Proceedings of the 2002 conference on Applications, technologies, architectures, and protocols for computer communications
Observation and analysis of BGP behavior under stress
Proceedings of the 2nd ACM SIGCOMM Workshop on Internet measurment
BGP routing stability of popular destinations
Proceedings of the 2nd ACM SIGCOMM Workshop on Internet measurment
Topology inference from BGP routing dynamics
Proceedings of the 2nd ACM SIGCOMM Workshop on Internet measurment
Identifying BGP routing table transfers
Proceedings of the 2005 ACM SIGCOMM workshop on Mining network data
Finding a needle in a haystack: pinpointing significant BGP routing changes in an IP network
NSDI'05 Proceedings of the 2nd conference on Symposium on Networked Systems Design & Implementation - Volume 2
Quantifying path exploration in the internet
IEEE/ACM Transactions on Networking (TON)
Impact of prefix-match changes on IP reachability
Proceedings of the 9th ACM SIGCOMM conference on Internet measurement conference
Understanding slow BGP routing table transfers
Proceedings of the 9th ACM SIGCOMM conference on Internet measurement conference
Longitudinal study of BGP monitor session failures
ACM SIGCOMM Computer Communication Review
BGP churn evolution: a perspective from the core
INFOCOM'10 Proceedings of the 29th conference on Information communications
Investigating occurrence of duplicate updates in BGP announcements
PAM'10 Proceedings of the 11th international conference on Passive and active measurement
The Journal of Supercomputing
Hi-index | 0.00 |
BGP routing updates collected by monitoring projects such as RouteViews and RIPE have been a vital source to our understanding of the global routing system. However the collected BGP data contains both the updates generated by actual route changes, and the updates of BGP routing table transfers resulted from BGP session resets between operational routers and the data collection stations. Since the latter is caused by measurement artifact, it is important to accurately separate out the latter from the former. In this paper, we present the design and evaluation of the minimum collection time (MCT) algorithm. Given a BGP update stream, MCT can identify the start and duration of each routing table transfer in the stream with high accuracy. We evaluated MCT performance by using three months of BGP data from all RIPE collectors. Our results show that out of the total 1664 BGP resets with 166 monitors, MCT can identify BGP routing table transfers with over 95% accuracy, and pinpoint the exact starting time of the detected table transfers in 83% of such cases. Accurate detection of BGP table transfers enables users to separate out real BGP routing changes and measurement artifacts, and can be used to measure and diagnose the BGP session failures.